. .
`
`.- .
`
`Molten Glass Corrosion Resistance of Immersed Combustion-Heating
`Tube Materials in E-Glass
`
`S. Kamakshi Sundaram,. Jen-Yen Hsu,' and Robert E Speyer.
`School of Materials Science and Engineering, Georgia Institute of Technology, Atlanta, Georgia
`30332-0245
`
`J Am. Cerum. SOC., 78 171 1940-46 (1995)
`
`The corrosion resistance of molybdenum, molybdenum
`disilicide, and a SiC(,,/AI,O, composite to molten E-glass at
`1550°C was studied. Mo showed no tendency to oxidize as it
`was immersed in soda-lime silicate glass in a parallel study.
`MoSi, was corroded by soluble molecular oxygen, leaving a
`Mo,Si, interface behind. The SiC(,,/AI,O, composite was
`corroded at a more rapid rate wherein the S i c component
`was oxidized to form amorphous silica and CO bubbles.
`Based on these results, the activity of soluble molecular
`oxygen in E-glass was determined to be in the range of 2.4 X
`10-14 to 2.0 x 10-8.
`
`I. Introduction
`HE work discussed herein is part of a larger research thrust
`
`T to develop materials for immersed gas-fired radiant burner
`
`tubes for glass melters. In the evaluation of candidate materials,
`our research has adopted a parallel approach, separately evalu-
`ating molten glass corrosion and combustion gas corrosion, as
`well as using thermodynamic simulations for both processes.
`Gas corrosion investigations are reported elsewhere.',' In the
`present paper, the corrosion resistance of candidate materials to
`a molten E-glass (fiberglass) composition is presented.
`Of the candidate materials evaluated, those presented here
`are molybdenum, molybdenum disilicide, and a SiC,,,/Al,O,
`composite ( p refers to particulate). Molybdenum has been used
`as the electrode material for electric melting and boosting in
`glass-melting tanks for many year^.^,^ Molybdenum disilicide
`has demonstrated a self-protective mechanism of forming an
`amorphous SiO, layer at the surface in an oxidizing atmo-
`sphere.' In the case of the SiC,,,/AI,O, composite, the combina-
`tion of high thermal conductivity of Sic and the refractory
`properties of A1,0, was anticipated to make it suitable for the
`combustion product side of the present application. This work
`closely parallels corrosion studies of these candidate materials
`in soda-lime-silicate glass at 1565°C.3 In that work, all three
`materials were found to corrode by oxidation reactions. An
`oxidation product interfacial region of MOO,(,, for Mo, and
`Mo,Si, for MoSi,, formed, which periodically cracked away
`due to interfacial stresses. CO/CO, bubbles formed as a reaction
`product in the corrosion of the SiC(,,/Al,O, composite.
`
`from MoSi, or SiC~,,/A1,03, alloys with platinum to destroy the
`crucible. Therefore, fusion-cast AZS (UNICORSOI , Corhart
`Refractories, Louisville, KY) crucibles with containment vol-
`umes of 3.81 cm diameter by 3.81 cm depth were used. AZS
`consists of 50.80 wt% A1,0,, 32.60 wt% ZrO,, 14.90 wt%
`SO,, 1.25 wt% Na,O, 0.13 wt% Fe,O,, and 0.12 wt% Ti0,.9
`Molybdenum (Johnson Matthey/AESAR, Ward Hill, MA)
`specimens were of 12.5 mm diameter by 12.5 mm length for
`the purpose of complete immersion. Chemical analysis for trace
`elements in as-received molybdenum showed less than 1 ppm
`of Al, Ca, Cr, Cu, Mg, Mn, Ni, Pb, Si, Sn, and Ti, less than
`2 ppm of C and 0, and less than 14 ppm of Fe. Cylindrical
`specimens of MoSi, (Kanthal Super 33, Kanthal, Bethel, CT)
`and SiC~,,/AI,O, (DuPont-LanxideTM Composites, Wilmington,
`DE) of 12.5 mm diameter by 60 mm length were used. As
`determined by scanning electron microscopy (SEM) image
`analysis, the as-received MoSi, contained 1.7 vol% Mo,Si,
`grains, generally in contact with amorphous aluminosilicate
`glassy pockets, which in turn comprised 18.6 vol% of the
`specimen.
`The E-glass (Owens/Corning Fiberglas, Granville, OH) used
`in the present investigation is a common fiberglass composition.
`A chemical analysis of the as-received glass, in weight percent,
`shows SiO, = 54.4%, CaO = 17.9%, A1,0, = 14.7%, B,O, =
`6.3%, MgO = 4.7%, Na,O = 0.6%, TiO, = 0.6%, F, = 0.5%,
`and Fe,O, = 0.33%.
`The as-received glass was crushed to -40 mesh, and the
`AZS crucible with 48.5 g of glass was placed inside an electri-
`cally heated furnace. The glass was heated to 1550°C and
`allowed to equilibrate at that temperature for 2 h to obtain
`bubble-free glass. Test specimens were then loaded into the
`crucible and exposed for 12, 24, or 48 h. Details of measure-
`ment and characterization procedures may be found in related
`published
`
`11. Experimental Procedure
`The ASTM standard method C621-84 for isothermal corro-
`sion resistance of refractories to molten glass was revised for
`the present in~estigation.~ Since molybdenum oxidizes and vol-
`atilizes above the glass line at temperatures exceeding -6OO"C,
`it was fully immersed. Soluble silicon in the melt, originating
`
`J. L. Smialek-contributing editor
`
`"L 0
`
`10-
`
`20
`30
`Time (hrs)
`
`40
`
`50
`
`Manuscript No. 194582. Received May 10,1993; approved February 16, 1995.
`Supported by the Gas Research Institute under Contract No. 5090-298-2073. Much
`of this work was performed at the New York State College of Ceramics at Alfred
`University.
`'Member, American Ceramic Society.
`
`Fig. 1. Corrosion of refractory containment materials at glass-line
`(solid lines) and half-down (dotted lines) regions. The error bars indi-
`cate the standard deviation in the measurement of specimen diameters
`before the corrosion tests.
`
`1940
`
`GE-1022.001
`
`

`

`July 1995
`
`Molten Glass Corrosion Resistance of Immersed Combustion-Heating Tube Materials in E-Glass
`
`1941
`
`Table I. Chemical Analyses (wt%) of Glass after Interaction with AZS at 1550°C
`Corrosion time (h) = 0
`24
`3
`6
`12
`36
`48
`54.4
`53.8
`54.5
`54.8
`53.5
`54
`53.2
`15.7
`15.6
`15.7
`17.4
`17.2
`18.1
`14.7
`0.15
`0.55
`1.01
`0
`0
`0
`0
`
`SiO,
`A1203
`ZrO,
`
`111. Results and Discussion
`The choice of A Z S was made after static corrosion testing
`of fusion-cast A Z S , MgO-partially-stabilized zirconia (PSZ)
`(Coors Ceramics, Golden, CO), fusion-cast chromia (Carborun-
`dum, Monofrax Refractories Division, Falconer, NY), and
`bonded chromia (Corhart Refractories, Buchannon, WV)
`refractories, following the ASTM C621-84 standard. The
`results of corrosion testing, presented in Fig. 1, indicate that
`AZS had an inferior corrosion resistance, both at the glass line
`and half-down regions, to bonded and fusion-cast chromia. The
`half-down corrosion rates of bonded chromia and PSZ could
`not be measured due to specimen cracking, for the latter,
`undoubtedly from destabilization of the cubic crystal structure.
`Fusion-cast chromia demonstrated slight recession up to 24 h,
`followed by an apparent increase in cross-sectional area after
`48 h of exposure. This remarkable feature consistently appeared
`during three separate trials, in both half-down and glass-line
`measurements. Two possibilities are suggested: (1) The high
`vapor pressure of chromia tended to expand porous regions and
`increased specimen dimensions by creep. (2) The chemistry of
`solid solutions within the refractory changed via incorporation
`of ions from the glass into the chromia structure, in turn causing
`a volume expansion. Bonded chromia did not show this dilation
`behavior. AZS refractory was, nevertheless, selected as the
`containment material over chromia refractories, to eliminate
`glass coloration.
`The chemical analyses of the glass after interaction with AZS
`for varying times are shown in Table I. No significant changes
`in these constituents were observed up to 12 h of exposure.
`Beyond 12 h, the A1,0, and ZrO, content increased; after 48 h
`of exposure, the A1,0, weight fraction increased to 2.4%, and
`ZrO, to 1.01%. The effect of this compositional change on the
`specimen corrosion rate is not known. Replenishing glass melt
`after every 12 h of exposure was considered and rejected, since
`it changes the present static testing to a semidynamic corrosion
`test.
`Figures 2(a and b) show the corrosion of cylindrical speci-
`mens by molten glass (in turn contained in AZS crucibles). In
`the case of glass-line corrosion, SiC,,,/Al,O, showed a corro-
`sion rate comparable to that of MoSi, for the first 24 h but
`accelerated beyond that. SiC,,J,/A1,O, corrosion appeared
`
`largely independent of positioning at the glass-line or half-
`down locations. MoSi, showed an order of magnitude decrease
`in recession rate at the half-down position as compared to the
`glass-line position. Immersed Mo showed the least corrosion,
`with no measurable corrosion for the first 24 h.
`(1) Molybdenum
`Figure 3 shows the Mwglass interfacial morphology; no
`interfacial layer can be seen. This is corroborated by a line EDS
`scan (not shown) in which a sharp decrease in Mo concentration
`with corresponding sharp increase in Si concentration was
`observed at the interface. XRD patterns starting from the glass
`side, penetrating into the bulk (not shown), did not show any
`interfacial phase at the Mo-glass interface.
`Neither MOO, nor MOO,,,, were detected at the interfacial
`region. For the equilibria,
`
`AGY,,,, = -268.6 kJ/mol, ao,,,,,,,, = 2.0 X
`Thus, the
`molecular oxygen activity in the molten glass would have to be
`greater than or equal to 2.0 X lo-' for this reaction to be
`favorable. The oxygen activity would have to equal or exceed
`7.4 X
`for MOO,,,, to be favorable.
`The concentration of molecular oxygen in alkali silicate
`melts has been shown" to be 1% of that in the atmosphere,
`which translates to a soluble molecular oxygen activity of -2 X
`lo-, for glass in an air atmosphere. This was consistent with an
`observed MOO, surface layer on Mo exposed to molten soda-
`lime-silicate glass,, since this oxygen activity was well above
`the critical value of 2.0 X lo-*. No information is available for
`the solubility of molecular oxygen in calcium aluminosilicates
`such as the E-glass composition studied herein. However, based
`on the fact that no molydenum oxides were observed, an oxy-
`gen activity below 2.0 X lo-' at 1550°C in E-glass is indicated,
`assuming that there were no kinetic hindrances to oxide
`formation.
`The enhancement in Mo corrosion between 24 and 48 h of
`exposure is suspected to be a result of solution of impurity
`phases from the AZS crucible. The presence of impurity phases
`such as As,O,, Na,SO, have been shown" to enhance the corro-
`sion of molybdenum in molten glass, but these phases are not
`present in AZS. However, solution of zirconia, iron oxide and
`
`C .z fa
`2 - 1
`
`.......................................
`
`.......' A t
`
`
`
`10
`
`40
`
`50
`
`"- 0
`
`10
`
`20
`30
`30
`20
`Time (hrs)
`Time (hrs)
`(a)
`(b)
`Fig. 2.
`(a) Specimen corrosion at the glass-line (dotted lines) and half-down regions (solid lines). (b) Magnified view of corrosion of half-down Mo
`and MoSi,. Propagated error calculations' result in error bars smaller than the symbol size and are therefore not shown. The dashed line shows radius
`reduction results corrected for the thickness of the original vitreous silica layer on as-received MoSi, (see text for discussion).
`
`40
`
`so
`
`GE-1022.002
`
`

`

`1942
`
`Journal of the American Ceramic Society-Sundaram et al.
`
`Vol. 78, No. 7
`
`Fig. 3. Molybdenum-glass interfacial region after 48 h of corrosion. M: molybdenum, G: glass. Left: secondary electron image. Right: backscat-
`tered image.
`
`titania, present in AZS, would be expected to alter the solubility
`of molecular oxygen in the glass. It is therefore possible that
`corrosion of the AZS crucible would cause the corrosion of Mo
`by increasing the solubility of molecular oxygen in E-glass.
`(2) Molybdenum Disilicide
`The as-received MoSi, had a fused silica surface layer of
`-10 pm thickness. After preheat treatment at 1000°C for
`10 min, the surface layer became somewhat thinner and discon-
`tinuous. If it is assumed that this vitreous coating mixed with
`the glass so that it was effectively removed, then the measured
`postcorrosion specimen radii would be affected, but not the
`trend in radius recession with exposure time. The dashed line
`in Fig. 2(b) shows the measured MoSi, specimen radii after
`subtraction of a 10 pm surface layer.
`Figure 4 shows XRD patterns starting from the glass, pene-
`trating through the interface, into the bulk. From the nearby
`glass to the interfacial region, molybdenum was identified. The
`major interfacial phase was identified to be Mo,Si,. Minor
`
`quantities of Mo,Si could not be confirmed or denied, since the
`most intense peaks of this phase overlap with those of Mo,Si,.
`An SEM EDS pattern was taken (not shown) of the surface
`from which the XRD second from the front in Fig. 4 was taken.
`Compared to as-received MoSi,, appreciable molybdenum
`enhancement relative to silicon was apparent, corresponding to
`an XRD pattern showing predominantly Mo,Si,.
`Figures 5(a and b) show the MoSi2-glass interfacial features
`at the glass-line and half-down regions, respectively. The
`interfacial layer had a thickness of 133.2 ? 4.3 pm at the half-
`down region, and 13.0 ? 3.2 pm at the glass-line region. From
`the backscattered images, the interfacial region was richer in
`atomically heavier atoms. Since oxygen is lighter than both
`metals, the interfacial region is not an oxide such as MOO, or
`SiO,. Mo being heavier than SI, the interfacial region is inter-
`preted to be Mo-rich. Further corroboration is provided by
`EDS line scans (Fig. 6). At the interface, the Si-to-Mo ratio
`is decreased. Although the Si intensity varies throughout the
`interfacial region, a general Si concentration gradient from
`
`8 0 6
`Fig. 4. XRD patterns as a function of penetration, analyzing glass, interfacial region, and molybdenum disilicide bulk. m: molybdenum, 1: MoSi,,
`and 2: Mo,Si,. XRD surfaces were prepared by repeatedly grinding a flat glass-exposed surface with 600 pm abrasive paper for 10 min and cleaning.
`
`GE-1022.003
`
`

`

`July 1995
`
`Molten Glass Corrosion Resistance of Immersed Conzbustion-Heating Tube Materials in E-Glass
`
`1943
`
`Molybdenum disilicide-glass interfacial region at (a) glass-line and (b) half-down regions after 48 h of corrosion. M: molybdenum
`Fig. 5.
`disilicide, I: interface, F: flaked-off layer, and G: glass. Left: secondary electron image. Right: backscattered image.
`
`interior to exterior is apparent. The fluctuations in Si and Mo
`concentrations in the MoSi, bulk, as well as the interfacial
`region, are attributed to dispersed pockets of amorphous alumi-
`nosilicate phase throughout the material. A small Mo EDS peak
`away from the interface suggests a limited molybdenum-rich
`waste stream into the glass. As implied from Fig. 5, at the glass
`line, the flaked-off reaction-product layer was drawn away from
`the interface by enhanced glass-line convection currents.” This
`corrosion product layer remained close to the interface at the
`half-down region. This also explains the decreased interfacial
`layer thickness at the glass line as compared to the half-down
`region.
`XRD, SEM, and EDS results provide conclusive evidence
`that the mechanism of MoSi, corrosion is removal of silicon
`from MoSi, to the glass, leaving a silicon-deficient molybde-
`num silicide interface. This is an oxidation reaction where oxy-
`gen was provided from soluble molecular oxygen: $MoSi, +
`= ?Mo,Si, + SiO,. Figure 7 is an Ellingham diagram
`02(g,ars)
`
`for equilibria of interest among compounds of Mo, Si, and 0.
`As implied by the lowest standard free-energy equilibria line,
`formation of silica from MoSi, via the most oxygen conserva-
`tive mechanism is favored-silica
`is formed by oxidation and
`removal of silicon from MoSi, to form Mo,Si,, rather than
`complete oxidation of the silicon component, the molybdenum
`component, or both. In addition, the most favored reaction is
`that most conservative in silicon removal from the compound;
`e.g., Mo,Si, formed rather than Mo,Si or Mo. If no MoSi, is
`locally available, then it would be thermodynamically favorable
`for Mo,Si, to convert to Mo,Si, and by the same argument, for
`Mo,Si to convert to Mo.
`These predictions are consistent with the above experimental
`results where, at the interface, Mo,Si, formed. This interface
`would not be expected to convert to Mo,Si and Mo if MoSi, is
`locally available, since the latter has a greater affinity for molec-
`ular oxygen. An appreciable volume change would be associ-
`ated with silicon removal from MoSi,, and it is suspected that
`
`GE-1022.004
`
`

`

`1944
`
`3000
`
`h % 2000
`8
`v .-
`F z
`g 1000
`
`U
`
`Journal of the American Ceramic SocietpSundaram et al.
`1500
`
`Vol. 78, No. 7
`
`1200
`
`h % 2 900
`Y
`.-
`F $ 600
`
`8-
`U
`
`300
`
`0
`0
`
`20
`
`40
`60
`Distance (pm)
`
`80
`
`100
`
`0
`0
`
`20
`
`40
`60
`Distance (pm)
`
`80
`
`100
`
`Fig. 6. Line EDS of molybdenum disilicide-glass interfacial region at (a) glass-line, (b) half-down regions.
`
`when a critical thickness has been reduced to Mo,Si,, it dis-
`lodges from the interface and floats away as debris. No longer
`near regions of MoSi,, local dissolved oxygen converts the
`Mo,Si, to Mo,Si, and Mo,Si to Mo, which was observed by
`XRD in regions in the glass extended away from the interface.
`The rate of corrosion was undoubtedly determined by the
`combined effects of the rates of coarsening of interfacial layers
`and the critical thicknesses at which interfacial stresses caused
`the coatings to crack and flake away, in turn, forcing new
`interfaces to form. Silicon diffusion through the Mo&
`is
`required to form vitreous silica, which dissolves into the glass.
`An appreciable volume percentage of silicon vacancies are
`expected to exist at the Mo,Si, interface; as silicon removal
`from the lattice continues, silicon vacancies form until a struc-
`tural collapse to a unit cell of different stoichiometry results.
`This is corroborated by the Si concentration gradiant shown by
`line EDS in Fig. 6. This type of interface would foster easy Si
`diffusion along these vacancy sites.
`
`(3) SI'C~,,~A~203
`Figures 8(a and b) show the SiC,,,/Al,O,-glass
`interfacial
`features at the glass-line and half-down regions, respectively.
`Bubbles formed within the glass, near the composite surface at
`both the glass-line and half-down regions. Near the glass line,
`the bubbles are seen in contact with the bulk of the specimen.
`CO or CO, can be the product of reaction between SIC and
`oxygen, depending on oxygen activity.', Competing oxidation
`reactions are considered in Fig. 9. The Sic portion of the
`composite oxidizes most favorably, to form silica and carbon
`monoxide. Since the CO-CO, equilibria fall below the Richard-
`son line (for Mo oxidation), the conversion of CO to CO, is not
`favorable, and the bubbles are expected to be purely carbon
`monoxide. Based on a standard Gibbs energy change of
`AGY,,,, = -534.67 kJ/mol, for the SiC,,,/CO,,, equilibria, and
`a CO bubble pressure of 1 atm, the molecular oxygen activity
`for the equilibria is uo, = 4.8 X
`Oxygen activities above
`this cause the reaction to be favorable.
`
`100
`
`0
`
`-100
`
`ka k
`-200
`a
`fi- W 0" -300
`
`a 3 -400
`3 g -500
`
`m 3 -600
`-700
`
`-800
`
`-900
`0
`
`.* 2 x 10-8
`
`1000
`2000
`Temperature (K)
`
`3000
`
`Fig. 7. Ellingham diagram of compounds of interest from MoSi,-oxygen interaction. A Richardson line of 2.0 X lo-' represents the highest
`possible molecular oxygen activity based on the results of Mo in E-glass. The dot-dashed line indicates the test temperature (1823 K). Thermody-
`namic data were taken from standard tab~lations.'~
`
`GE-1022.005
`
`

`

`July 1995
`
`Molten Glass Corrosion Resistance of Immersed Combustion-Heating Tube Materials in E-Glass
`
`1945
`
`SiC~,,/Al,O,-glass interfacial region after 48 h of corrosion at (a) glass-line and (b) half-down regions. C : composite, E: epoxy, and G: glass.
`Fig. 8.
`Left: secondary electron image. Right: backscattered image.
`
`The presence of bubbles in contact with a refractory in mol-
`ten glass has been demonstrated to be a mechanism of highly
`accelerated corrosion referred to as upward or downward
`“drilling,”’2 where convection driven by surface tension gradi-
`ents sweeps fresh glass to the interface and corrosion products
`away. Thus, oxygen-containing fresh glass would replace more
`viscous silica-rich glass. This corresponds well to the observed
`high corrosion rates for these specimens.
`(4) Remarks
`Oxidation did not occur with Mo immersed in E-glass, but
`did occur for MoSi, in the same glass. Mo,Si, debris converted
`to Mo,Si, which in turn converted to Mo. This could occur only
`for a soluble molecular oxygen activity of 2.4 X
`or higher.
`Thus, the oxygen activity of this composition of E-glass is
`and 2 X lo-’. This is an appre-
`bracketed between 2.4 X
`ciably lower oxygen activity than implied by corrosion studies
`in soda-lime-silicate glasses, where a soluble oxygen activity
`was based on the formation of a MOO, scale
`above 2.0 X
`on Mo.
`Immersed MoSi, corroded more slowly in E-glass than in
`soda-lime-silica glass, while the opposite was true for Sic,,,/
`
`A1,0,. It is interpreted that, with the lower oxygen activity in
`E-glass, the driving force for oxidation of MoSi, would be
`lessened. The 15°C higher temperature used in soda-lime-
`silicate glass studies may have also been a factor. The faster
`corrosion of SiC~,,/Al,O, in E-glass is attributed to the lower
`In a Iower
`viscosity of E-glass in this temperature
`viscosity glass, bubbles can more easily nucleate at the glass-
`solid interface and rise away from the interface. Removal of
`gaseous bubbles from the interface would keep product gases
`from suppressing the oxidation reaction.
`Extrapolation of these corrosion results to predicting the
`lifetime of components made of these materials in commercial
`glass tanks is mitigated by conditions unlike those of the experi-
`ment. So long as immersed Mo is not exposed to an E-glass
`which has random composition variances which would in turn
`permit increased molecular oxygen solubility, it should be
`indefinitely stable. For MoSi,, the convective flow forces com-
`mon in a glass tank may act to accelerate removal of the Mo,Si,
`interfacial layer and accelerate its corrosion behavior as com-
`pared to that presently reported. The atmosphere over the glass
`under combustion heating would be predominantly nitrogen,
`
`GE-1022.006
`
`

`

`1946
`
`Journal of the American Cei
`vamic Societ-undaram
`
`et al.
`
`Vol. 78, No. 7
`
`cracked away, fostering the formation of a new interfacial layer
`and continued specimen recession. SiC(,,/Al,O, receded at a
`rate approximately one order of magnitude greater, where the
`formation of CO bubbles fostered a mechanism of enhanced
`corrosion.
`
`References
`‘W. Lin, J. Y. Hsu, S. K. Sundaram, and R. F. Speyer, “Reactions between
`Flue Gas and Ceramic Composite Tubes for Radiant Heating”; presented at
`the 93rd Annual Meeting and Exposition of the American Ceramic Society,
`Cincinnati, OH, April 28-May 2, 1991 (Abstract No. 19-SII-91).
`*J. Y. Hsu, S. K. Sundaram, W. Lin, and R. F. Speyer, “Phase Equilibria
`Simulation between Immersed Radiant Tube Compositions and Molten Glass”;
`presented at the 93rd Annual Meeting and Exposition of the American Ceramic
`Society, Cincinnati, OH, April 28-May 2, 1991 (Abstract No. 38-SII-91).
`’S. K. Sundaram, J. Y. Hsu, and R. F. Speyer, “Molten Glass Corrosion
`Resistance of Immersed Combustion-Heating Tube Materials in Soda-Lime
`Silicate Glass,”J. Am. Ceram. Soc., 77 [6] 1613-23 (1994).
`*ASTM C621-84, “Standard Test Method for Isothermal Corrosion Resistance
`of Refractories to Molten Glass.” American Society for Testing and Materials,
`Philadelphia, PA, 1992.
`’L. Penberthy, “Electric Melting of Glass”; pp. 114-19 in Electric Melfing in
`the Glass Industry. Compiled by A. G. Pincus and G. M. Diken. Books for
`Industry and The Glass Industry Magazine, New York, 1976.
`‘R. Eck, “The State-of-the-Art of Molybdenum Fabrication,” Ceram. Eng.
`Sci. Proc., 6 [3-4] 274-86 (1985).
`’V. Bizzam, B. Linder, and N. Lindskog, “Molybdenum Disilicide Heating
`Elements: Meeting Advanced Ceramics Requirements,” Am. Cerum. Soc. Bull.,
`68 [lo] 1834-35 (1989).
`*S. K. Sundaram and R. F, Speyer, “Evaluation of Molten Glass Corrosion for
`Highly Corrosion Resistance Materials,” J. Test. Evul., in press.
`9Product Literature on Unicor501 AZS Refractory; p. Al. Corhart Refracto-
`ries, Louisville, KY, 1991.
`‘“R. H. Doremus, Glass Science; pp. 121-24. Wiley, New York, 1973.
`“K. Ooka, “Corrosion of Molybdenum by Molten Glasses,” J. Cerum. Assoc.
`Jpn., 72 [71 110-15 (1964).
`12V. L. Burdick, “The Corrosive Nature of Molten Glass”; pp. 545-61 in
`Introduction to Glass Science. Edited by L. D. Pye, H. J. Stevens, and W. C.
`Lacourse. Plenum Press, New York, 1972.
`‘’K. E. Spear, R. E. Tressler, Z. Zheng, and H. Du, “Oxidation of Silicon
`Carbide Single Crystals and CVD Silicon Nitride,” .I. Am. Cerum. Soc., 69 [9]
`674-81 (1986).
`’‘I. Barin, Thermochemical Data of Pure Substances, Part I and II; pp. 139,
`140, 157, 158, 292, 293, 301, 309, 929, 931-33, 1347-50, 1433-35, 1534-37,
`1687,1688, 1690. VCH, Weinheim, Germany, 1989.
`”E. J. Homyak, Owens-Illinois; private communication, 1990.
`I‘D. Miller, Owens-Coming Fiberglas; private communication, 1990.
`
`-700 ’
`
`0
`
`I
`I
`lo00
`2000
`Temperature (K)
`
`I
`3000
`
`Fig. 9. Ellingham diagram used to justify the chemistry of bubbles
`formed in the glass with oxidation of SiC,,,/Al,O,.
`
`water vapor, and carbon dioxide. This atmosphere would form
`less-oxidizing soluble species in the glass and may therefore
`allow MoSi, and SiC,,,/Al,O, to be stable in E-glass. With
`these caveats, the recession rates of MoSi, and SiC,,,/Al,O,
`components in E-glass, based on 24 h of corrosion, are 1.7 and
`38.6 cm/year, respectively.
`
`IV. Concf usion
`Mo showed no oxide layer formation when immersed in
`E-glass at 1550°C in an air atmosphere. MoSi, and SiC~,,/Al,O,
`showed reaction products indicating reactions with soluble
`molecular oxygen. A Mo,Si, layer formed which periodically
`
`GE-1022.007
`
`